Sentence examples for model we trained from inspiring English sources

When training the character-based model, we trained the character-based model both from the left to the right and from the right to the left in each sentence.

To compare the performance of our model in recovering positive hits in RNAi screens with the canonical supervised learning model, we trained a Random Forest model with the same prior features and training sets used in our model.

To assess the out-of-sample predictive accuracy of the KBBN SITVIT model we trained the model on SITVIT-Train and tested it on SITVIT-Test.

For each initial model, we trained our algorithm on five different training datasets and then went on to test its performance on a test dataset.

For the single-domain model, we trained on one single PDZ domain associated with all the interaction data, and tested for the held-out PDZ domain.

To further test a parallel memory trace model, we trained flies with 0.3% DEET and 1 M sucrose, a combination with which no immediate odor avoidance or approach performance is evident, and blocked either the rewarding or aversive dopaminergic neurons during training.

To gather a library of diverse models, we trained a total of 282 GBM, RF and SVM models.

All the models we trained had two input nodes: one for pretrained word-level embeddings and another one for encoded token strings.

To better understand the influence of the experimental error in GP modeling, we trained 15 models for each dataset with increasing levels of noise with both the radial and the normalized polynomial (NP) kernel, thus leading to a total number of 90 models.

Finally, in addition to exploring holdout validation performance of the ensemble models, we trained and tested the ensemble models on different gold standards.

As for the acoustic modeling, we train the HMMs of INITIAL and FINAL units, which corresponds to the semi-syllables in Mandarin Chinese.